Page 11 of 13
Re: turbine modelling
Posted: Thu Jul 12, 2012 1:10 pm
by JamesMcNaughton
imrgra is an integer so remove the .d0. If you type "code_saturne compile" whilst in the SRC folder it will tell you any issues which might save you time waiting for answers on here.
The subroutines you use are up to you. But I believe usclim and usini1 are required yes. You might also want to try initialising the flow using usiniv (there are examples in it).
Re: turbine modelling
Posted: Thu Jul 12, 2012 4:10 pm
by stage75
It's OK for the (GRADRC) I have teh convergence for this GRADRC , using (imrgra=3 as you see James) but I have always this error which I don't undestand its source !
Code: Select all
/home/scc/Téléchargements/tmp/installer/ncs-2.0.4/src/base/cs_sles.c:1746: Erreur fatale.
Jacobi: error (divergence) solving for Turb Dis
and
Code: Select all
/home/itouche/Téléchargements/tmp/installer/ncs-2.0.4/src/base/cs_sles.c:1746: Erreur fatale.
Jacobi: error (divergence) solving for Turb Kin
Re: turbine modelling
Posted: Fri Jul 13, 2012 8:29 am
by JamesMcNaughton
It's the same error as your previous posts! The calculation is diverging. Did you try the things I said previously?
Re: turbine modelling
Posted: Fri Jul 13, 2012 8:54 am
by stage75
Yes James I try your advice, I write (Imrgra = 3) after the step-time, that's why I don't have the mesage of ((NON CONVERGENCE DE GRADRC) that I had previously
Re: turbine modelling
Posted: Fri Jul 13, 2012 9:51 am
by JamesMcNaughton
I mean here
http://code-saturne.org/forum/viewtopic ... t=80#p5991 - it is the same error yes?
Did you try the advice I gave in the following posts?
Re: turbine modelling
Posted: Fri Jul 13, 2012 10:13 am
by stage75
Yes, I checked the mesh and it's good (no error detected in the meshes).
For CFL I have the default parameters (I haven't changer it)
Code: Select all
! turbulent viscosity
ipp = ipppro(ipproc(ivisct(iphas)))
nomvar(ipp) = 'Turb Viscosity'
ichrvr(ipp) = 1
ilisvr(ipp) = 1
ihisvr(ipp,1) = -1
! Courant number
ipp = ipppro(ipproc(icour(iphas)))
nomvar(ipp) = 'CFL'
ichrvr(ipp) = 1
ilisvr(ipp) = 0
ihisvr(ipp,1) = -1
! Fourier number
ipp = ipppro(ipproc(ifour(iphas)))
nomvar(ipp) = 'Fourier Number'
ichrvr(ipp) = 1
ilisvr(ipp) = 0
ihisvr(ipp,1) = -1
Re: turbine modelling
Posted: Fri Jul 13, 2012 4:04 pm
by Jacques Fontaine
stage75 wrote:Yes, I checked the mesh and it's good (no error detected in the meshes).
[...]
Hello,
There is never any error (or rarely) during the check_mesh. You must open the check_mesh log check criteria histograms and/or visualize the check_mesh result in ParaView to determine the mesh quality.
Regards,
JF
Re: turbine modelling
Posted: Sun Jul 15, 2012 9:19 pm
by stage75
Hi,
Finally I will recomncer from the beginning and I will not taken into account the solid parties (turbine , tank) as you said James, I'll take only the fluid portion and I'll see what I would have the same error!
I started the simulation from 0s until 0.1s and I want to continue the simulation until 1s how I could do starting from the solution at 0.1 to continue my simulation until 1s!?
Thanks
Re: turbine modelling
Posted: Mon Jul 16, 2012 11:55 am
by stage75
Hello all,
I put icorio = 0 instead of 1 and I get no error up to 2s. But from 2s I get an error "jacobi" divergence "rel velo"??
How could I do to boost my simulation without leaving 0s but atarting from 2s and changing the time step that I had to 0s to 2s? maybe I'll get the convergence in this manner.!!
Thanks
Re: turbine modelling
Posted: Mon Jul 16, 2012 1:28 pm
by stage75
Please there is someone who knows? because every time I need to start soon t = 0s!
Thank's